sign in sign up
<<
Home , Lithosphere, Transition zone (Earth)

and Letters 236 (2005) 249–257 www.elsevier.com/locate/epsl

The – boundary in the North-West Atlantic region

P. Kumar a,b, R. Kind a,c,*, W. Hanka a, K. Wylegalla a, Ch. Reigber a, X. Yuan a, I. Woelbern a, P. Schwintzer a,1, K. Fleming a, T. Dahl-Jensen d, T.B. Larsen d, J. Schweitzer e, K. Priestley f, O. Gudmundsson g, D. Wolf a

aGeoForschungsZentrum Potsdam, Telegrafenberg, 14473 Potsdam, Germany bNational Geophysical Research Institute, Uppal Road, Hyderabad-500007, India cFreie Universita¨t Berlin, Malteserstr. 74-100, 12249 Berlin, Germany dGeological Survey of Denmark, Oster Voldgade 10, 1350 Copenhagen K, Denmark eNORSAR, Institutsveien 25, 2027-Kjeller, Norway fUniv Cambridge, Bullard Lab, Madingley Rise, Cambridge, CB3 0EZ, United Kingdom gDanish Lithosphere Center, Oster Voldgade 10, 1350 Copenhagen K, Denmark Received 30 November 2004; received in revised form 21 April 2005; accepted 23 May 2005 Available online 1 July 2005 Editor: V. Courtillot

Abstract

A detailed knowledge of the thickness of the lithosphere in the north Atlantic is an important parameter for understanding plate in that region. We achieve this goal with as yet unprecedented detail using the seismic technique of S-receiver functions. Clear positive signals from the – boundary and negative signals from a mantle discontinuity beneath Greenland, Iceland and Jan Mayen are observed. According to seismological practice, we call the negative phase the lithosphere–asthenosphere boundary (LAB). The seismic lithosphere under most of the Iceland and large parts of central Greenland is about 80 km thick. This depth in Iceland is in disagreement with estimates of the thickness of the elastic lithosphere (10–20 km) found from postglacial rebound data. In the region of flood basalts in eastern Greenland, which overlies the proposed Iceland plume track, the lithosphere is only 70 km thick, about 10 km less than in Iceland which is located directly above the proposed plume. At the western Greenland coast, the lithosphere

* Corresponding author. GeoForschungsZentrum Potsdam, Telegrafenberg, 14473 Potsdam, Germany Tel.: +49 331 288 1240; fax: +49 331 288 1277. E-mail addresses: [email protected] (P. Kumar), [email protected] (R. Kind), [email protected] (W. Hanka), [email protected] (K. Wylegalla), [email protected] (X. Yuan), [email protected] (I. Woelbern), [email protected] (K. Fleming), [email protected] (T. Dahl-Jensen), [email protected] (T.B. Larsen), [email protected] (J. Schweitzer), [email protected] (K. Priestley), [email protected] (O. Gudmundsson), [email protected] (D. Wolf). 1 Has passed away on Dec. 2004.

0012-821X/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.epsl.2005.05.029 250 P. Kumar et al. / Earth and Planetary Science Letters 236 (2005) 249–257 thickens to 100–120 km, with no indication of the Iceland plume track identified. Below Jan Mayen the lithospheric thickness varies between 40 and 60 km. D 2005 Elsevier B.V. All rights reserved.

Keywords: Lithosphere–asthenosphere boundary; Iceland; Greenland; S-receiver functions

1. Introduction and adds a few more processing steps. Such steps are, as in the P-receiver function technique, source equal- High-viscosity lithospheric plates moving over a isation by deconvolution and distance move-out cor- lower-viscosity asthenosphere is a basic element of rection. Both steps are applied in order to enable the . Lithosphere and asthenosphere are summation of events from different distances and with originally mechanical definitions with regards to different magnitudes and source-time functions. This their reaction to forces acting over thousands or technique works very well, enabling observations of millions of years [1]. However, additional usages of the LAB with a resolution so far only known for the the term dlithosphereT have been introduced since Moho. then: thermal, seismic or chemical [2]. Obtaining In this work we determine the lithospheric thick- high-resolution seismic observations of the litho- ness for Iceland, Greenland and the island of Jan sphere–asthenosphere boundary (LAB) is not an Mayen. Greenland is a continent of Precambrian age easy task. Observations of low seismic velocities in (see [9] for a discussion of Greenland geology). the are interpreted as being indicative of Darbyshire et al. [10] observed a thickening of the the asthenosphere. The first seismic observations of an lithosphere from 120 to 200 km going from east to dasthenospheric channelT were obtained by Gutenberg west in southern Greenland using the surface wave [3] at about 100 km depth. Therefore, seismologists technique. sometimes call the LAB the dGutenberg discontinuityT Iceland is thought to be one of the classic mantle (mostly in oceanic regions). The lithosphere is seis- plumes [11] interacting with a mid-ocean ridge, al- mologically divided into two parts, the crust and the though this view is disputed [12]. The crustal thick- mantle lithosphere, the latter being the high-velocity ness (up to 45 km) is several times thicker than that lid on top of the asthenosphere. However, high-reso- expected for (e.g. [13,14]). White and lution seismic body-wave observations of the LAB are McKenzie [15] concluded that the large thickness of very rare. This is in contrast to the Moho, which is the Icelandic crust is a result of magmatic intrusions. globally a much better documented discontinuity. So The mechanical lithosphere of Iceland is thought to be far, most information about the thickness of the lith- very thin (10–20 km), judging from rapid postglacial osphere comes from low-resolution surface-wave uplift [16]. A 10–20 km thick lithosphere would mean observations (e.g. [4]). The thickness of the litho- that the lower crust and Moho are located within the sphere is considered to be close to zero at mid- asthenosphere. There are also arguments that east ocean ridges, about 200 km beneath stable , Iceland may be a continental splinter [17–19]. Seismic with 80–100 km being the global average. Thybo and surface-wave studies have furthermore found indica- Perchuc [5] suggest the existence of a global zone of tions of a 50–110 km thick lithosphere under parts of reduced velocity at about 100 km depth underlying Iceland [18,20,21]. Evans and Sacks [22] found be- continental regions, based on controlled source seis- tween Iceland and Jan Mayen a lithospheric thickness mic data. Li et al. [6] and Kumar et al. [7] obtained of 50 km from surface wave data, typical for young detailed maps of the LAB around the Hawaiian island oceans. Vinnik et al. [23] have shown a negative chain and in the Tien Shan–Karakoram region, respec- discontinuity at a depth of 80 km beneath all of Ice- tively, using the S-receiver function technique [8]. land, using essentially the same data and technique This new technique complements the traditional S to that we have used here. Jan Mayen is a small volcanic P conversion method (applied to S or SKS phases) island located about 600 km north of Iceland and P. Kumar et al. / Earth and Planetary Science Letters 236 (2005) 249–257 251 about 400 km east of Greenland on the Jan Mayen a fracture zone. It is considered to be a micro continent BRE [24]. HOT15 HOT09 66û KLU HOT12 HOT13 HOT29 HOT08 HOT07 HOT16 HOT06 ASBS HOT14 SKOT HOT04 HOT27 HOT17 HOT05 HOT18 BLOL HOT26 HOT25 2. Data and observations HOT03 BIRH HOT02 NYD ASKJ HOT24 HOT19 BORG HOT23 For the Greenland study, we used seismic data HOT30 HOFF LJOP KAF from the GLATIS and NEAT experiments [25,26]. 64û HOT21 During these experiments, seismic stations were HOT22 deployed along the Greenland coast and on the ice sheet for periods ranging from several months to -25û -20û -15û several years. For Iceland, we used the publicly avail- able data of the ICEMELT and experi- ments [27,28]. In addition, we used data from 68û b permanent IRIS [29] and GEOFON [30] stations, 4 7 and from two seismic stations on the island of Jan 1 Mayen: JMI, operated by the Norwegian National Seismic Network (http://www.Ifjf.uib.no/Seismologi/ 66û nnsn/nsninfo2.html) and JMIC, operated by NOR- 2 5 8 SAR [31]. The locations of the stations and of the S-P piercing points at 80 km depth are shown in 64û Figs. 1–3. Since converted phases are usually weak 3 6 signals, a number of records must be summed to obtain a good signal-to-noise ratio. We defined non- overlapping regions on Greenland, Iceland and Jan 62û -25û -20û -15û -10û

Fig. 2. a) Location of the temporary seismic stations of the HOT- SPOT and ICEMELT experiments and of the IRIS station BORG. b) Location of the piercing points of the S receiver functions at 80 km 80û I depth. Also marked are the regions (1–8) used for the summation of II ALE the seismic traces. E C DAG Mayen, where all traces with piercing points in 70û DBG UPN NGR these regions have been summed to form one record A F SUM HJO that is representative for the entire region (see Figs. GDH IS2 1–3). The number of traces stacked within each 60û SFJ SOE region are more than 20 (see Figs. 1, 2, and 3a). NUK DY2 D ANG Individual traces (not summed traces) are shown in B G 0û PAA Fig. 3b for the Jan Mayen stations as an example of the quality of our data. The individual S-receiver -60û -20û functions and its stacks are also shown from the -40û region 2 in Iceland (Fig. 4a), and from the region C Fig. 1. Location of the temporary and permanent seismic stations of in Greenland (Fig. 4b). the GLATIS and NEAT experiments, and of IRIS and GEOFON in Fig. 5a shows the summation traces of the P Greenland (and one in NE Canada, ALE) used in this study (re- component for all regions. Zero time is the S-arrival versed triangles), and of the piercing points of the S-receiver func- tions at 80 km depth (plus signs). The regions used for the time, where negative times indicate the period prior to summation of seismic traces have also been marked and are labelled the S arrival. We have rotated the ZNE components A–F and I, II. into the RTZ system using the theoretical backazi- 252 P. Kumar et al. / Earth and Planetary Science Letters 236 (2005) 249–257

-10

û û 0

û û

-30 -20 a

68 û 9A 71.5û

64û

71û JMIC JMI

9B 70.5û -10û -9û -8û -7û

Backazimuth (deg.) 8 22 24 24 25 26 32 63 68 75 138 237 340 b 0

-5 LAB

-10 Time (s)

-15 18 24 24 25 26 29 33 63 75 86 138 278 353 9A 9B

Fig. 3. a) Location of the seismic stations JMI and JMIC of the Norwegian National Seismic Network (JMI) and of NRSAR (JMIC) on the island of Jan Mayen and of the piercing points of the S-receiver functions used. Also marked are two regions, 9A and 9B, that were used for the summation of seismic traces. b) Individual (not summed) S receiver functions observed at the Jan Mayen stations. The LAB is clearly observed in all traces. Events with backazimuths of 8–258(region 9A) show the LAB earlier then the events arriving from all other azimuths (region 9B).

muth angle. R and Z are then rotated a second time a SUM into the P-SV system. The angle of incidence is 0 defined by the minimum of energy in the P compo- Moho -5 nent at the arrival time of the S phase. All traces are -10 LAB distance moveout corrected before summation, using

Time (s) -15 a reference slowness of 6.4 s/deg based on the -20 IASP91 global reference model [32]. A bandpass filter of 4–20 s has been applied. Two precursor b SUM 0 phases are clearly visible in Fig. 5a, the Moho and Moho -5 a second phase, which we term the LAB. The arrival LAB -10 times of the Moho and LAB must be measured at the

Time (s) -15 maximum (or minimum) of the signal due to the deconvolution. The arrival times of the LAB in sec- -20 onds may be multiplied by a factor of 10 (according Fig. 4. The individual S-receiver functions from Iceland and Green- to the IASP91 model) to obtain the LAB depth esti- land and the corresponding stacked traces on the left. (a) is from the mate in kilometers. The possible sources of errors in region 2 in Iceland (Fig. 2) and (b) is from the region C in Green- land (Fig. 1). These two plots clearly show the presence of coherent depth determination are primarily from the time to negative phases at around 8 to 9 s. The small scatter in the data depth conversion (due to the uncertainty in lithospher- (about 0.5 s) are a source of errors in depth determination. ic velocity) and the selection of the times of the P. Kumar et al. / Earth and Planetary Science Letters 236 (2005) 249–257 253 a BIIA3DCEG2568174FI9B9A in Fig. 5a is especially remarkable, since phases in the 0 MOHO uppermost mantle are difficult to observe at a high -5 resolution using other techniques. In P-receiver func- tions, the time window of the LAB arrival is heavily -10 LAB disturbed by crustal reverberations. These reverbera- Time (s) -15 tions are not present in the S-receiver functions, because the converted phases are S precursors, where- -20 as multiples arrive after the main phase. The Moho is b 0 usually well observed in P-receiver functions, which MOHO have shorter periods and therefore higher resolution. -5 Hence, we will concentrate here on the LAB phase. It -10 should be noted that stations on the Greenland Ice LAB

Time (s) Sheet (regions C and D) produce nearly undisturbed -15 S-receiver functions in contrast to P-receiver func- -20 tions, which are heavily disturbed by reverberations c in the ice layer [25].InFig. 5b,c we have modelled 0 the observed data presented in Fig. 5a using theoret- MOHO ical seismograms (Fig. 5b), the models used are dis- 50 played in Fig. 5c. Complete theoretical seismograms are computed (plane-wave Haskell-matrix formalism) 100 LAB using simple models consisting of a homogeneous Depth (km) crust on top of a homogeneous mantle lithosphere, 150 both overlaying a homogeneous asthenosphere. Fig. 5. a) S-receiver functions in the NW Atlantic obtained at However, we have not modelled the sharpness of seismic stations on Greenland, Iceland and Jan Mayen. Zero is the LAB. Only the depths of the Moho and of the the arrival time of the S phase. Negative times indicate the period LAB have been adjusted to fit the different arrival prior to the S-signal arrival. Characters on top of the figure indicate times of both phases in different regions. We believe the regions used for the summation of the seismic traces. The this simple modelling, which is sufficient to repro- location of these regions is shown in Fig. 2. Two seismic phases are marked: the crust–mantle boundary (Moho) and the lithosphere– duce all features of the S precursors (in S-receiver asthenosphere boundary (LAB). The traces are sorted by the arrival functions), provides evidence for the interpretation of times of the LAB phase. The time scale is valid for a move out the observed phase called LAB as the lithosphere– correction slowness of 6.4 s/degree. Multiplication of the LAB asthenosphere boundary. times by a factor of 10 gives approximately the LAB depth. b) A map of the LAB depth, as determined from the Theoretical receiver functions of the S velocity models in c). Moho and LAB are the only two phases predicted by the model (no data, is shown in Fig. 6. The LAB is 80 km deep multiples). The velocities are kept fixed in all models in c) (Vs under most of Iceland and large parts of Greenland crust=3.58 km/s, Vs lid=4.5 km/s, Vs astheno- (yellow in Fig. 6). In Iceland, we have only two sphere=3.9 km/s). Only the depth of both the discontinuities are regions where the LAB differs from 80 km depth, varied to fit the travel times. region 3 (90 km, light blue) and region 4 (70 km, light red). Along the west coast of Greenland, the phases due to scattering (~0.5 s) in the data. Here, we LAB is at 100–120 km depth (dark blue). North of estimated the maximum error bounds due to the Greenland, in the transition to the Arctic Ocean, the various uncertainties as being less than about 10 km LAB shallows from 120 km (dark blue) to 70 km for the depth estimation. The signs of the Moho and (light red). The shallowest LAB on Greenland is the LAB are opposite, indicating downward increas- observed in region F (70 km, light red). The Jan ing and decreasing velocity jumps, respectively. The Mayen LAB varies between 40 and 60 km depth traces in Fig. 5a are sorted according to the arrival (dark red) and is clearly the shallowest observed in times of the LAB. Characters on the top correspond to this study. Details of the Jan Mayen LAB are given the regions (see Figs. 1–3). The clearness of the LAB in Fig. 3b. A surprising result is that the entire 254 P. Kumar et al. / Earth and Planetary Science Letters 236 (2005) 249–257

km 19 120 limit of 1Â10 Pa s for the underlying viscosity. Sigmundsson and Einarsson [33], from an assessment 80 I 100 û D II e of lake-tilt data resulting from the melting of the 80 p Greenland t Vatnajoekull Ice Cap, inferred a sub-lithosphere vis- E h 18 19 60 cosity of between 1Â10 À5Â10 Pa s. More re- 70 C û cently, Thoma and Wolf [34] used present-day 9A,B 40 changes in uplift and gravity to propose thicknesses A Jan Mayen F of 10 to 20 km for the lithosphere, and a viscosity 1 4 7 16 18 D range of 7Â10 À3Â10 Pa s for the asthenosphere 60 2 5 8 û G viscosity. Similar values are given by Sjoeberg et al. 3 6 Iceland B 0û [35], based on GPS campaign results from around Vatnajoekull. The result of a 50–100 km thick seismic -60 û lithosphere beneath Iceland, as seen in seismic sur- û -20 -40û face-wave data and in S-receiver function data, there- fore appears to conflict with these estimates of the Fig. 6. Bathymetric map of the NW Atlantic with regions marked elastic lithosphere thickness (10–20 km). To examine where the depth of the lithosphere was obtained by the S-receiver functions. Location of seismic stations and of piercing points of S how we may reconcile the two estimates, Fig. 7 pre- receiver functions at the LAB are shown in Figs. 1–3. sents predictions of vertical uplift rates around Vatna- joekull based on a range of earth models that

Icelandic LAB has practically the same depth as a large parts of central Greenland. Depth (km) Model 1 Model 2 Model 3Model 4 Model 5 Model 6 Model 7 Model 8 0 10 η 18 20 Low-viscosity channel lvc = 10 Pa s 3. Discussion and conclusions Elastic 30 Similar to our results, Li et al. [6] and Kumar et al. 40 [7] have also observed the LAB in Hawaii and in the 50 60 Tien Shan–Karakoram area, respectively. These find- 70 ings lead us to suppose that the LAB may be a 80 η 18 globally existing and observable discontinuity, com- Low-viscosity zone lvz = 10 Pa s parable to the Moho. The S-receiver function tech- b nique therefore appears to be a very useful tool for )

-1 30 mapping the global LAB. Model 1 Model 5 Model 2 Model 6 25 Model 3 Model 7 However, as mentioned above, there are different Model 4 Model 8 usages of the phrase dlithosphereT. In , a 20 Sjöpberg et al. (2004) velocity decrease marks the lower boundary of the 15 lithosphere, whereas the mechanical thickness of the 10 lithosphere may be estimated from other data, such as 5 postglacial rebound observations. In Iceland, these 0 data sets give different results. Glacial-isostatic ad- justment (GIA) studies infer that the viscosity strati- -5 fication underlying Iceland consists of a thin elastic -10 0 20 40 60 80 100 120 140 Vertical-displacement rate (mm yr lithosphere overlying a low-viscosity asthenosphere. Distance from ice-cap centre (km) For example, Sigmundsson [16], using post-glacial Fig. 7. a) The earth models used to examine the possibility that the -level observations and assuming a 10 km elastic 80 km discontinuity represents a lower elastic layer bounding a low- lithosphere (the approximate maximum depth of viscosity channel. b) The predicted and measured [38] vertical uplift in SW-Iceland), estimated an upper rates from the centre of the Vatnajoekull Ice Cap. P. Kumar et al. / Earth and Planetary Science Letters 236 (2005) 249–257 255 incorporate a low-viscosity channel within the seismic The suggested track of the Iceland plume in Green- lithosphere, as well as earth-model end members with land is marked by tertiary flood basalts in our regions lithosphere thicknesses of 80 and 10 km. Although a F and A (Fig. 6, [9,39]). The lithosphere in region F is detailed parameter space study was not carried out, the thinnest for Greenland (70 km) while region A, our assumed viscosity (1018 Pa s) is of a similar order which is located at the west coast, shows a lithosphere of magnitude as values found from other studies (e.g. thickness of about 120 km (Fig. 6). This could mean 0.3–2Â1019 Pa s [36], 1992; 3Â1018 À3Â1019 Pa s that the track of the Iceland plume is traceable to the [37]). These results are compared with recent GPS- east coast of Greenland, but not to the central part or based up-lift rates [38]. Model 6, with a channel the west coast, where the lithosphere has possibly had thickness of 50 km, gives a reasonable fit to the enough time to regain its normal thickness. This could observations. However, an alternative explanation of also mean that the Iceland plume has caused, in a this discrepancy may be that the seismic lithosphere, manner similar to the Hawaii plume, a delayed reju- reacting to short term (seconds or minutes) elastic venation of the lithosphere (starting from our ob- forces, is not related to the elastic lithosphere reacting served 80 km lithosphere at Iceland) when the plate to longer term (years to thousands of years) forces. passed over it. In Fig. 8, the residual geoid signal in the NW-Atlantic is shown, obtained from a combina- tion of the most recent CHAMP [40]and GRACE [41] satellite data with aerogravimetry and terrestrial grav- a ity data from the Arctic Gravity Project (ArcGP, [42]) after all geoid features larger than 2500 km were filtered out. We note positive residual geoid heights over the as well as in our region F at 70 the east coast of Greenland. There is no continuation of these positive residual heights across Greenland to 40 the flood basalt region at the west coast (region A). 0 The larger positive anomaly in southern Greenland remains unexplained. The thinning of the lithosphere in eastern Greenland (region F in Fig. 8b) is clearly visible. The thinning occurs in the mantle part of the lithosphere, not in the crust, because the Moho in -10 -5 0510m region F is not updoming.

b ADC F 1 5 8 0 MOHO Acknowledgements -5

-10 This research was supported by the GFZ Potsdam, LAB Time (s) the Deutsche Forschungsgemeinschaft (DFG) and the -15 European Commission (EC). PK was supported by a -20 stipend from the Deutsche Akademische Austausch- dienst (DAAD). PK is also grateful to the Director, Fig. 8. a) Residual geoid signal in the NW Atlantic region, obtained from a combination of CHAMP and GRACE satellite and aero- NGRI and DG, CSIR, India, for granting him leave to gravimetry/terrestrial gravity data (after removal of all geoid com- carry out this research work. Some of the seismic ponents N2500 km). The suggested trace of the Iceland plume is stations were provided by the station pool of the marked (dashed line, [43]). The numbers give the estimated time of GFZ Potsdam, data is stored in the GEOFON archive. the plume location in millions years BP. The white encircled regions We thank Joachim Saul, Franz Barthelmes and Neil near the plume track mark Tertiary flood basalts in eastern and western Greenland. b) S-receiver functions from Fig. 5a aligned Mcglashan for their support. R.K.’s research visit at along the plume trace. The characters on the top of the figure mark NORSAR was supported by the EC Programme Ac- the regions described in Figs. 1, 2, and 3a. cess to Research Infrastructures (Contract HPRI-CT- 256 P. Kumar et al. / Earth and Planetary Science Letters 236 (2005) 249–257

2002-00189). We also wish to thank Bob Trumbull for [16] F. Sigmundsson, Post-glacial rebound and asthenosphere discussions. Reviews by two anonymous reviewers viscosity in Iceland, Geophys. Res. Lett. 18 (1991) 1131–1134. and G. R. Foulger improved the manuscripts. The [17] H.E.F. Amundsen, et al., in: B. Jamtveit, H.E.F. Amundsen computations have been done in SeismicHandler (K. (Eds.), 15th Kongsberg Seminar, Norway, Kongsberg, 2002. Stammler). [18] I.T. Bjarnason, The Lithosphere and Asthenosphere of the Iceland Hot Spot, IUGG, 2003. [19] T. Fedorova, W.R. Jacoby, H. Wallner, Crust–mantle transition and Moho model for Iceland and surroundings from seismic, References topography, and gravity data, 396 (2005) 119–140. [1] J. Barrell, The strength of the Earth’s crust, J. Geol. 22 (1914) [20] I.T. Bjarnason, How thick is the lithosphere below Iceland? 680ff. How large is the velocity inversion and what does it mean? [2] D.L. Anderson, Lithosphere, asthenosphere, and perisphere, (Abstract), EOS Trans. AGU 80 (46) (1999) F645 (Fall Meet. Rev. Geophys. 33 (1995) 125–149. Suppl.). [3] B. Gutenberg, Untersuchungen zur Frage bis zu welcher Tiefe [21] A. Li, R.S. Detrick, Structure of crust and upper mantle die Erde kristallin ist, Z. Geophys. 2 (1926) 24–29. beneath Iceland from Rayleigh wave tomography (Abstract), [4] J. Dormann, M. Ewing, J. Oliver, Study of shear-wave velocity EOS Trans. AGU 84 (46) (2003) F645 (Fall Meet. Suppl.). distribution in the upper mantle by mantle Rayleigh waves, [22] J.R. Evans, I.S. Sacks, Deep structure of the Iceland plateau, Bull. Seismol. Soc. Am. 50 (1960) 87–115. J. Geophys. Res. 84 (1979) 6859–6866. [5] H. Thybo, E. Perchuc, The seismic 88 discontinuity and [23] L.P. Vinnik, G.R. Foulger, Z. Du, Seismic boundaries in the in the continental mantle, Science 275 mantle beneath Iceland: a new constraint on temperature, (1997) 1626–1629. Geophys. J. Int. 160 (2005) 533–538. [6] X. Li, R. Kind, X. Yuan, I. Woelbern, W. Hanka, Rejuvenation [24] S. Kodaira, R. Mjelde, K. Gunnarsson, H. Shiobara, H. of the lithosphere by the Hawaiian plume, 427 (2004) Shimamura, Structure of Jan Mayen microcontinent and 827–829. implication for its evolution, Geophys. J. Int. 132 (1998) [7] P. Kumar, X. Yuan, R. Kind, G. Kosarev, The lithosphere– 383–400. asthenosphere boundary in the Tien Shan–Karakoram region [25] T. Dahl-Jensen, T.B. Larsen, I. Woelbern, T. Bach, W. Hanka, from S receiver functions: evidence for continental , R. Kind, S. Gregersen, K. Mosegaard, P. Voss, O. Gudmunds- Geophys. J. Lett. 32 (2005), doi:10.1029/2004GL022291. son, Depth to Moho in Greenland: receiver function analysis [8] V. Farra, L.P. Vinnik, Upper mantle stratification by P and S suggests two proterozoic blocks in Greenland, Earth Planet. receiver functions, Geophys. J. Int. 141 (2000) 699–712. Sci. Lett. 205 (2003) 379–393. [9] N. Henriksen, A.K. Higgins, F. Kalsbeek, T.C.R. Pulvertaft, [26] S. Pilidou, K. Priestley, E. Debayle, O. Gudmundsson, Ray- Greenland from archean to quarternary, Geol. Greenl. Surv. leigh wave tomography in the North Atlantic: high resolution Bull. 185 (2000). images of the Iceland, Azores and Eifel mantle plumes, [10] F.A. Darbyshire, T.B. Larsen, K. Mosegaard, T. Dahl-Jensen, LITHOS 79 (2005) 453–474. O. Gudmundsson, T. Bach, S. Gregersen, H. Pedersen, W. [27] I.T. Bjarnason, C.J. Wolfe, S.C. Solomon, G. Gudmundson, Hanka, A first detailed look at the Greenland lithosphere and Initial results from the ICEMELT experiment: body-wave upper mantle, using Rayleigh wave tomography, Geophys. J. delay times and shear-wave splitting across Iceland, Geophys. Int. 158 (2004) 267–286. Res. Lett. 23 (1996) 459–462, doi:10.1029/96GL00420. [11] W.J. Morgan, Convection plumes in the , Nature [28] G.R. Foulger, M.J. Pritchard, B.R. Julian, J.R. Evans, R.M. 230 (1971) 42–43. Allen, G. Nolet, W.J. Morgan, B.H. Bergsson, P. Erlendsson, [12] D.L. Anderson, Thermal state of the upper mantle; no role for S. Jakobsdottir, S. Ragnarsson, R. Stefansson, K. Vogfjord, , Geophys. Res. Lett. 27 (2000) 3623–3626. The seismic anomaly beneath Iceland extends down to the [13] F.A. Darbyshire, I.T. Bjarnason, R.S. White, O.G. Flovenz, mantle and no deeper, Geophys. J. Int. 142 Crustal structure above the Iceland mantle plume imaged by (2000) F1–F5. the ICEMELT refraction profile, Geophys. J. Int. 135 (1998) [29] R. Butler, T. Lay, K. Creager, P. Earl, K. Fischer, J. Gaherty, G. 1131–1149. Laske, B. Leith, J. Park, M. Ritzwoller, J. Tromp, L. Wen, The [14] M.A. Allen, G. Nolet, W.J. Morgan, K. Vogfjord, M. Nettles, global seismographic network surpasses its design goals, EOS G. Ekstrom, B.H. Bergsson, P. Erlendsson, G.R. Foulger, S. Trans. AGU 80 (23) (2004) 225–229. Jakobsdottir, B.R. Julian, M. Pritchard, S. Ragnarson, R. [30] W. Hanka, A. Heinloo, K.-H. Jaeckel, Networked seismo- Stefanson, Plume driven plumbing and crustal formation in graphs: GEOPHONE real time data distribution, ORFEUS Iceland, J. Geophys. Res. 107 (B8) (2002) 2163, doi:10.1029/ Electron. Newsl. 2 (2000) 24 (http://orfeus.knmi.nl/newsletter/ 2001JB000584. vol2no3/heofon.html). [15] R. White, D. McKenzie, Magmatism at zones: the gener- [31] J. Fyen, K. Iranpour, Near real time data at NORSAR for ation of volcanic continental margins and flood basalts, J. CTBT monitoring, ORFEUS Electron. Newsl. 5 (2) (2003) Geophys. Res. 94 (1989) 7685–7729. (http://orfeus.knmi.nl/newsletter/vol5no2/norsar.html). P. Kumar et al. / Earth and Planetary Science Letters 236 (2005) 249–257 257

[32] B.L.N. Kennett, E.R. Engdahl, Traveltimes for global earth- [39] M. Storey, A.K. Pedersen, O. Stecher, S. Bernstein, H.C. quake location and phase identification, Geophys. J. Int. 105 Larsen, L.M. Larsen, J.A. Baker, R.A. Duncan, Long-lived (1991) 429–465. postbreakup magmatism along the east Greenland margin: [33] F. Sigmundsson, P. Einarsson, Glacial-isostatic crustal move- evidence from shallow-mantle metasomatism by the Iceland ments caused by historical volume change of the Vatnajoekull plume, Geology 32 (2004) 173–176, doi:10.1130/G19889.1. ice cap Iceland, Geophys. Res. Lett. 19 (1992) 2123–2126. [40] C. Reigber, H. Jochmann, J. Wu¨nsch, S. Petrovic, P. Schwint- [34] M. Thoma, D. Wolf, Inverting uplift near Vatnajoekull, Ice- zer, F. Barthelmes, K.-H. Neumayer, R. Ko¨nig, C. Fo¨rste, G. land, in terms of lithosphere thickness and viscosity stratifica- Balmino, R. Biancale, J.-M. Lemoine, S. Loyer, F. Perosanz, tion, in: M. Sideris (Ed.), Gravity, Geoid and , Earth gravity field and seasonal variability from CHAMP, in: Springer-Verlag, Berlin, 2001, pp. 97–102. C. Reigber, H. Lu¨hr, P. Schwintzer, J. Wickert (Eds.), Earth [35] L. Sjoeberg, M. Pan, E. Asenjo, S. Erlingsson, Glacial rebound Observation with CHAMP—Results from Three Years in near Vatnajoekull, Iceland, studied by GPS campaigns in 1992 Orbit, Springer-Verlag, Berlin, 2004, pp. 25–30. and 1996, J. Geodyn. 29 (2000) 63–70. [41] C. Reigber, R. Schmidt, F. Flechtner, R. Ko¨nig, U. Meyer, [36] G.R. Foulger, C.-H. Jahn, G. Seeber, P. Einarsson, B.R. Julian, K.-H. Neumayer, P. Schwintzer, S.Y. Zhu, An Earth gravity K. Heki, Post-rifting stress relaxation at the divergent plate field model complete to degree and order 150 from GRACE: boundary in Iceland, Nature 358 (1992) 488–490. EIGEN-GRACE02S, J. Geodyn. 39 (2005) 1–10. [37] F.F. Pollitz, I.S. Sachs, Viscosity structure beneath northeast [42] R.S. Forsberg, Kenyon, Gravity and geoid in the Arctic Iceland, J. Geophys. Res. 101 (1996) 17771–17779. region—The northern gap now filled, Proc. of 2nd GOCE [38] L.E. Sjoeberg, M. Pan, S. Erlingsson, E. Asenjo, K. Arna- User Workshop (on CD-ROM), ESA SP-569, ESA Publica- son, Land uplift near Vatnajoekull, Iceland, as observed by tion Division, 2004. GPS in 1992, 1996 and 1999, Geophys. J. Int. 159 (2004) [43] L.A. Lawver, R.D. Mu¨ller, Iceland hotspot track, Geology 22 943–948. (1994) 311–314.